Scenario-driven dynamic analysis for comprehending large software systems

Understanding large software systems is simplified when a combination of techniques for static and dynamic analysis is employed. Effective dynamic analysis requires that execution traces be generated by executing scenarios that are representative of the system's typical usage. This paper presents an approach that uses dynamic analysis to extract views of a software system at different levels, namely (1) use cases views, (2) module interaction views, and (3) class interaction views. The proposed views can be used to help maintainers locate features to be changed. The proposed approach is evaluated against a large software system, the Mozilla Web browser